Dual Signaling Functions of the Hybrid Sensor Kinase RpfC of Xanthomonas campestris Involve Either Phosphorelay or Receiver Domain-Protein Interaction*

The hybrid sensor kinase RpfC positively regulates the expression of a range of virulent genes and negatively modulates the synthesis of the quorum sensing signal diffusible signal factor (DSF) in Xanthomonas campestris. Three conserved amino acid residues of RpfC implicated in phosphorelay (His198 in the histidine kinase domain, Asp512 in the receiver domain, and His657 in the histidine phosphotransfer domain) were essential for activation of the production of extracellular enzymes and extracellular polysaccharide (EPS) virulence factors but were not essential for repression of DSF biosynthesis. Domain deletion and subsequent in trans expression analysis revealed that the receiver domain of RpfC alone was sufficient to repress DSF overproduction in an rpfC deletion mutant. Further deletion and alanine scanning mutagenesis analyses identified a peptide of 107 amino acids and three amino acid residues (Gln496, Glu504, and Ile552) involved in modulating DSF production. Co-immunoprecipitation and far Western blot analyses suggested an interaction between the receiver domain and RpfF, the enzyme involved in DSF biosynthesis. These data support a model in which RpfC modulates two different functions (virulence factor synthesis and DSF synthesis) by utilization of a conserved phosphorelay system and a novel domain-specific protein-protein interaction mechanism, respectively. This latter mechanism represents an added dimension to conventional two-component signaling paradigms.

Two-component regulation is the predominant form of signal recognition and response coupling mechanism used by bacteria to sense and respond to diverse environmental stresses and cues ranging from common environmental stimuli to host signals recognized by pathogens and bacterial cell-cell communication signals (1)(2)(3). Bacteria chromosomes encode numerous two-component systems, implying diversified roles in signal modulation of microbial physiology and ecology. For example, the human pathogen Pseudomonas aeruginosa and plant pathogen Xanthomonas campestris pv. campestris carry, respectively, Ͼ60 and Ͼ100 pairs of genes encoding two-com-ponent systems (4,5). Typically, recognition of a signal by the sensor component results in autophosphorylation at a histidine residue; the phosphoryl group is subsequently transferred to an aspartate residue in the CheY-like receiver (REC) 2 domain of the cognate response regulator (2,6).
The structures of both sensors and regulators are modular, and numerous variations in domain architecture and composition have evolved to tailor to specific needs in signal perception and signal transduction (2,7). Among the extremely diversified family of histidine kinase sensors, the simplest (also known as orthodox kinases) consists of only sensing and kinase domains. The more complex hybrid sensors contain, in addition to sensing and kinase domains, a REC domain typical of two-component regulators and in some cases a C-terminal histidine phosphotransferase (HPT) domain (2,6). The family of sensor kinases with this latter domain organization includes ArcB of Escherichia coli, BvgS of Bordetella sp., GacS of Pseudomonas sp., and RpfC of X. campestris pv. campestris (8 -13). In the case of such hybrid sensor-regulator kinases, the phosphoryl group from the autophosphorylated histidine (His 1 ) residue is transferred to an aspartate (Asp 1 ) residue of the REC domain and is further relayed to a histidine residue (His 2 ) in the HPT domain. Subsequently, the His 2 transfers the phosphoryl group to an aspartate (Asp 2 ) residue in the REC domain of the cognate response regulator (12)(13)(14). Although the REC domain of hybrid sensor kinases is implicated in phosphorelay, it is by no means clear that this is its sole function (15).
In X. campestris pv. campestris, the hybrid sensor kinase RpfC and cognate regulator RpfG are implicated in the positive regulation of biofilm dispersal and the production of virulence factors (16,17). This two-component system is believed to respond to the cell-cell communication signal DSF (16 -19), which has been characterized as cis-11-methyl-2-dodecenoic acid (20). Synthesis of DSF requires an enzyme encoded by the rpfF gene (18). Recent microarray and genetic analyses have revealed that DSF also modulates additional functions associated with stress resistance and adaptation (21). A number of lines of evidence support a role for RpfC/RpfG in the perception and transduction of the DSF signal. The addition of DSF can restore virulence factor production and induce biofilm dispersal in rpfF mutants but not in rpfC and rpfG mutants, respec-tively (16,17). Furthermore, the RpfC/RpfG two-component system has been reconstructed in P. aeruginosa and shown to confer responsiveness to exogenously added DSF, as seen through its effects on swarming motility (19). Importantly, mutation of rpfC (but not of rpfG) leads to overproduction of DSF. These findings suggest that the RpfC sensor kinase may control two signaling pathways in X. campestris pv. campestris, with one activating virulence factor production and the other inhibiting DSF biosynthesis. The former pathway is dependent on the RpfG response regulator, but the latter apparently is not. The work in this paper had the aim of establishing the molecular mechanisms underlying this dual signaling action of RpfC.
As outlined above, RpfC is a hybrid sensor kinase in which the sensory input and kinase domains are fused to a receiver domain and a C-terminal HPT domain. By domain deletion and site-directed mutagenesis approaches, we have shown here that RpfC transduces signals for the regulation of virulence factor production through a phosphorelay system but modulates DSF biosynthesis through a domain-specific protein-protein interaction mechanism involving the REC domain. This latter finding, which offers an insight into possible additional roles of the REC domain of hybrid sensor kinases, presents a new dimension to the conventional two-component signaling paradigms.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Growth Conditions-The wild-type X. campestris pv. campestris strain Xc1 has been described previously (20). The Xc1 derivatives carrying various mutations or constructs were described in Figs. 1-5. X. campestris pv. campestris strains were grown at 30°C in YEB medium (22), unless otherwise stated. E. coli strains were grown at 37°C in LB medium. Antibiotics were added at the following concentrations when required: 100 g/ml kanamycin, 50 g/ml rifampicin, and 15 g/ml tetracycline; X-gluc (5-bromo-4-chloro-3indolyl ␤-D-glucopyranoside) was included in the medium at 60 g/ml for the detection of GUS (␤-glucuronidase) activity. DSF signal was added to the medium in a final concentration of 3 M when necessary.
Chromosomal Deletions in RpfC and Preparation of Constructs for in Trans Expression-RpfC was analyzed using the Simple Modular Architecture Research Tool (SMART). The open reading frame and the coding sequences for various domains of RpfC were deleted using the allelic exchange vector pK18mobsacB following the methods described previously (17). Briefly, for generation of the rpfC deletion mutant Xc1⌬C, two rpfC DNA fragments rpfC-1 (the 5Ј region of ϳ531 bp) and rpfC-2 (the 3Ј region of ϳ649 bp) were amplified using two primer pairs, rpfC-1-FOR and rpfC-1-REV and rpfC-2-FOR and rpfC-2-REV (supplemental Table S1). The resultant DNA fragments were cleaved with BamHI and ligated by T4 DNA ligase. The fusion fragment rpfC-12 was then amplified using the ligation mixture as the template with primer pair rpfC-1-FOR and rpfC-2-REV. The fusion fragment was cloned into the SmaI site of the vector pK18mobsacB. After sequence verification, the recombinant plasmid was mobilized into strain Xc1 by triparental mating. Transconjugants were selected on LB medium supplemented with rifampicin and kanamycin. A second selection was done on LB medium containing 5% (w/v) sucrose and rifampicin to select for resolution of the vector by a second crossover event. The in-frame deletion of rpfC was confirmed by PCR and sequencing. Similar methods were applied to generate the HPT domain deletion mutant RpfC⌬1, the REC-HPT domain deletion mutant RpfC⌬2, the HK-REC-HPT domain deletion mutant RpfC⌬3, and the transmembrane-HK-REC domain deletion mutant RpfC⌬4, using the primers listed (supplemental Table S1).
For the preparation of in trans expression constructs, the coding sequence of the REC domain of RpfC or its truncated versions was obtained by PCR amplification using the primers listed in supplemental Table S1. The PCR fragments were cleaved with BamHI and HindIII and cloned under the control of the lac promoter in the broad host range vector pLAFR3. The recombinant constructs were sequence-verified and mobilized into the RpfC null mutant Xc1⌬C by triparental mating. The resultant transformants were selected on LB medium supplemented with rifampicin and tetracycline.
Quantification of DSF and Virulence Factor Production-DSF synthesis, bioassay, and quantification were performed as described previously (20). DSF signals were extracted from the supernatants of bacterial cultures 40 h after inoculation, unless otherwise indicated. For determination of biofilm formation, 1 ml of bacterial cell culture at A 600 ϭ 1.6 was centrifuged at 10,000 rpm for 2 min, and the existence of a gum-like substance on the top of the bacterial pellet was checked as described previously (21). Quantification of EPS production and the activities of extracellular enzymes were performed as described previously (21,23,24).
RNA Extraction and Reverse Transcription (RT)-PCR Analysis-The detailed methods for RNA extraction and oligomicroarray analysis have been described previously (21). Briefly, bacterial cells at A 600 ϭ 1.6 were harvested by centrifugation at 4°C for 4 min at 10,000 rpm. RNA was purified by using an RNeasy midicolumn (Qiagen) following the protocol provided by the manufacturer. RT-PCR analysis was done using the Qiagen OneStep RT-PCR kit following the manufacturer's instructions. The primers used for RT-PCR analysis were listed in supplemental Table S1, and a total of 250 ng of total RNAs were used for each reaction.
In Situ Site-directed and Alanine Scanning Mutagenesis-Three conserved amino acid residues of RpfC predicted to be involved in phosphorelay (His 1 198 in the HK domain, Asp 1 512 in the REC domain, and His 2 657 in the HPT domain) were identified via the sequence alignment with the following homologues: RpfA (NCBI accession number(s) U62023), LemA (M80477), BvgS (M25401), GacS (AB219364), ArcB (X53315), and CheY (M13463). The three conserved residues were changed to alanine or valine by using substituted PCR primers (supplemental Table S1). The resultant PCR fragments were cloned into the SmaI site of pK18mobsacB. The recombinant constructs were verified by DNA sequencing and mobilized into strain Xc1 by triparental mating. Transconjugants were selected on LB medium supplemented with rifampicin and kanamycin. The second selection was done on YEB medium containing 5% (w/v) sucrose and rifampicin. The potential mutants were selected based on DSF production and biofilm formation phenotypes.
The point mutation was verified by PCR amplification of the corresponding DNA fragment and DNA sequencing.
For alanine scanning mutagenesis of the REC domain, the coding region was amplified using the primers rpfC-F3 and rpfC-R3 listed in supplemental Table S1 and cloned into the vector pGEMT-easy. Point mutation was conducted using the QuikChange site-directed mutagenesis kit following the manufacturer's instructions. After DNA sequencing verification, the mutated REC fragments were cut by BamHI and HindIII and cloned under the control of the lac promoter in expression vector pLAFR3. These constructs were then separately mobilized into strain Xc1⌬C by triparental mating.
Anti-FLAG Co-immunoprecipitation-The REC coding sequence was fused in-frame by PCR with that of FLAG using two primers listed in supplemental Table S1. After digestion with BamHI and HindIII, the PCR fragment was cloned in vector pLAFR3. The construct was transferred to the rpfC deletion mutant Xc1⌬C by conjugation. The expression of FLAG-REC fusion protein was confirmed by Western blot analysis and DSF bioassay. A total soluble protein sample was prepared when bacterial cell density reached 1.0 at A 600 and then applied onto EZview TM Red Anti-FLAG M2 affinity gel (Sigma) following the manufacturer's instructions. In brief, the gel was washed with TBS buffer (50 mM Tris HCl, 150 mM NaCl, pH 7.4), and the FLAG-tagged protein and its binding proteins were eluted with 0.1 M glycine HCl, pH 3.5. After condensation with Microcon YM-10 (Amicon), the eluted proteins were resolved by SDS-PAGE and stained with Coomassie Blue. Visible protein bands were excised from the gel, and the peptide sequences were deciphered by mass spectrometry (quadrupole time-of-flight).
Protein Purification and Anti-serum Preparation-The REC coding sequence was fused in-frame to the coding sequence of the His 6 tag in expression vector pET-14b (Novagen) and transformed into E. coli strain BL21 (DE3). Cells were grown at 28°C with shaking at 250 rpm to 0.7 at A 600 , isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 0.25 mM, and growth was continued overnight at 18°C with a gentle shaking at 200 rpm. The cells were harvested by centrifugation at 4000 rpm for 30 min and resuspended in the lysis buffer (pH 8.0) containing 50 mM NaH 2 PO 4 , 0.3 M NaCl, 10 mM imidazole, and 0.1 mM protease inhibitor mixture (Sigma). The cells were sonicated on ice with 5 ϫ 15-s bursts and 90-s cooling intervals. The cell debris was removed by centrifugation at 14,000 rpm for 30 min. The supernatant was then filtered using a 0.45-l filter before adding to an affinity column containing Ni 2ϩ -chelating Sepharose fast flow resin (Amersham Biosciences) for affinity binding. The column was washed with a buffer solution of the same pH containing 50 mM NaH 2 PO 4 , 0.3 M NaCl, and 20 mM imidazole. The bound His 6 -REC protein was eluted from the column with a 250 mM imidazole gradient and used as an antigen to obtain polyclonal antisera by immunizing rabbits through subcutaneous injections at two-week intervals. RpfF and green fluorescent protein were also purified in the same way for far Western blot analysis. Preparation of recombinant AlbD protein was described previously (25).

Western Blot and Far Western Protein-Protein Interaction
Assay-Western blotting was performed as described previously (26). In vitro far Western blot assay was performed following the method of Hall (27). Briefly, purified REC (0.1 g), RpfF (10 g), AlbD (10 g), and green fluorescent (10 g) proteins were resolved by SDS-PAGE and transferred to an Immun-Blot TM polyvinylidene difluoride membrane (Bio-Rad), which was then blocked with phosphate-buffered saline containing 0.05% Tween 20 and 3% nonfat powdered milk overnight at room temperature. The blocked membrane was overlaid with REC domain protein (30 g/ml in blocking buffer) for 4 h at room temperature. After washing with phosphate-buffered saline with 0.05% Tween 20 four times, the blots were incubated with primary polyclonal anti-REC serum followed by washing and incubation with secondary goat anti-rabbit IgG(HϩL)-horseradish peroxidase conjugate (Bio-Rad). The hybridization signal was detected using SuperSignal West Pico chemiluminescent substrate (Pierce).

Conserved Phosphorelay Mechanism of RpfC Is Essential for
Induction of Virulence Gene Expression but Not for Down-regulation of DSF Biosynthesis-RpfC is a hybrid sensor regulator, consisting of a transmembrane, an HK, a REC, and a HPT domain (Fig. 1A). Multiple sequence alignment analysis of RpfC and homologues identified several potential residues implicated in phosphotransfer, including the autophosphorylatable His 1 198 in the HK domain, the phosphor-accepting Asp residue (Asp 1 512 ) in the REC domain, and the phosphor-accepting His 2 657 in the HPT domain of RpfC ( Fig. 1A and supplemental Fig. S1). Similarly, we also identified the potential phosphoraccepting aspartate (Asp 2 80 ) residue in the REC domain of response regulator RpfG (data not shown). The conserved structure implies that the RpfC/RpfG two-component system may sense and transduce signals (including DSF) through the His 1 3 Asp 1 3 His 2 3 Asp 2 multiple phosphorelay system.
To determine whether the conserved phosphorelay mechanism in RpfC is involved in regulation of the dual functions, i.e. induction of virulence factor production and down-regulation of DSF production, we substituted His 1 198 and His 2 657 of RpfC with alanine and Asp 1 512 with valine by site-directed mutagenesis. These altered rpfC alleles were transferred to the chromosome to replace the wild type. RT-PCR analysis showed that these point mutations did not affect the expression of rpfC (supplemental Fig. S2) but resulted in decreased production of EPS and reduced activity of cellulase and protease, which was similar to the RpfC null mutant Xc1⌬C (Fig. 1B). Surprisingly, however, strains expressing the RpfC phosphotransfer-deficient variants H198A, D512V, and H657A produced a low level of DSF similar to the parental wild-type strain Xc1 (Fig. 1B) and unlike the rpfC deletion mutant Xc1⌬C, which produced elevated levels of DSF. These findings suggested that the conserved His 198 -Asp 512 -His 657 phosphorelay mechanism is required for induction of extracellular enzyme and EPS virulence factors but not for the RpfC-dependent inhibition of DSF synthesis.

The REC (but Not the HPT) Domain Is Required for Repression of DSF Biosynthesis-To address the mechanistic basis of
RpfC modulation of DSF biosynthesis, we first generated a panel of mutants with in-frame deletions in the chromosomal copy of the rpfC gene, which expressed variants of RpfC lacking one or more domains (Fig. 2A). RT-PCR analysis confirmed that these deletions did not affect the expression of the altered rpfC protein (supplemental Fig. S2). As expected, deletion of any of the three domains (HK, REC, or HPT) containing the conserved phosphorelay residues resulted in decreased activity of cellulase (Fig. 2B) and reduced EPS production and protease activity (data not shown). However, although deletion of both REC and HPT domains led to overproduction of DSF, the truncated RpfC lacking only the HPT domain remained active in the repression of DSF biosynthesis (Fig. 2C). Furthermore, all of the deletion mutants (RpfC⌬2, RpfC⌬3, and RpfC⌬4) lacking the REC domain overproduced the signal (Fig. 2, A  and C). These data, which seem to exclude the involvement of HPT in modulation of DSF signal generation, suggest a role for the REC domain in this function.
The Isolated REC Domain Can Repress DSF Biosynthesis-To further test the hypothesis that the REC domain is involved in RpfC-dependent repression of DSF biosynthesis, we cloned the coding region of this domain (the amino acid residues 450 -599 of RpfC) under the control of the lac promoter in the expression vector pLAFR3 for in trans expression in the rpfC deletion mutant Xc1⌬C and the wild-type strain Xc1. Fig. 3 shows that Xc1⌬C produces an elevated level of DSF, whereas overexpression of the REC domain in Xc1⌬C reduced DSF production to a  level less than the wild-type control. Consistent with these findings, expression of REC in wild-type strain Xc1 decreased DSF production to an undetectable level (Fig. 3).
The REC domain contains 150 amino acids. To locate the minimum active region, we generated seven further deletion variants of the REC domain (Fig. 4A). Variants R⌬1-R⌬4 lack the first 15, 29, 45, 60 amino acids, and R⌬5-R⌬7 lack the last 15, 29, and 45 amino acids of the REC domain, respectively. The constructs encoding these truncated REC variants were introduced into Xc1⌬C, and the effect of deletion on the repression of DSF biosynthesis (Fig. 4A) and the level of expression of each of the truncated proteins (Fig. 4B) were determined by Western analysis. These combined analyses suggested that the N-terminal amino acids 1-29 and the 15 amino acids at the C-terminal of REC are not essential for the repression of DSF biosynthesis. Western analysis failed to detect any REC derivatives with further deletion at either the N or C terminus. The findings were extended by in trans expression of the truncated REC domain without the first 29 N-terminal amino acids and the last 15 C-terminal amino acids (Fig. 4A, R⌬8). When expressed in Xc1⌬C, this peptide of 107 amino acids could repress DSF biosynthesis to a level lower than that in the wild-type strain Xc1, throughout growth (Fig. 4C).
Identification of Key Amino Acid Residues Implicated in REC Down-regulation of DSF Biosynthesis-The minimal REC region for repression of DSF biosynthesis contains 107 amino acids, including 47 hydrophobic amino acids (Fig. 5A). For identification of the key amino acid residues involved in downregulation of DSF biosynthesis, we employed alanine-scanning mutagenesis to alter each of the 90 amino acids (with the exception of the 17 alanines). These modified peptides were expressed in trans in mutant Xc1⌬C, and DSF production was determined. The analysis led to the identification of three amino acids (Gln 496 , Glu 504 , and Ile 552 ) for which alteration to an alanine residue reduced but did not totally abolish the REC repressor activity (Fig. 5B). To test for any synergistic effect among these three amino acid residue alterations, two constructs for in trans expression of the REC domain with double and triple alterations (Q496A,I552A and Q496A,E504A,I552A, respectively) were generated. However, no significant difference in DSF production was noticed when comparing the   Fig. 3), and DSF was bioassayed when bacterial cell density reached ϳA 600 ϭ 2.4. B, effect of deletion on protein expression. The total soluble protein extracts from the above strains were separated by electrophoresis and hybridized using RECspecific antiserum. C, DSF production profiles of wild-type strain Xc1, the rpfC deletion mutant Xc1⌬C, and Xc1⌬C expressing the minimal active region of REC (shown as R⌬8 in A).
effects of the expression of these multiply altered proteins with the corresponding single alterations (Fig. 5B).
Identification of RpfF as a REC-interacting Protein by Coimmunoprecipitation and Far Western Analysis-On the basis of the above findings, we posited that the effect of RpfC on DSF synthesis was due to binding of a (regulatory) protein by the REC domain and that such binding may be influenced by the RpfC conformation. To test the first part of this hypothesis, we searched for REC-interacting proteins using co-immunoprecipitation. A construct expressing the REC domain tagged with a 9-amino-acid FLAG peptide (see "Experimental Procedures") was introduced into the rpfC deletion mutant. The total proteins isolated from the cell culture at A 600 ϭ 1.0 were passed through the EZview TM Red Anti-FLAG M2 affinity gel (Sigma) for affinity binding. Electrophoretic separation of the proteins eluted from the affinity column and subsequent analysis by mass spectrometry identified four proteins in addition to the bait FLAG-REC as elongation factor Tu-B, RpfF, 2-hydroxyhepta-2,4-diene-1, 7-dioateisomerase/5-carboxymethyl-2-oxohex-3-ene-1, 7-dioatedecarboxylase, and 30 S ribosomal protein S3 (Fig. 6). The finding that RpfF, an enzyme essential for DSF biosynthesis, was able to bind to the REC domain was intriguing. To test whether any of the other three interacting proteins were also required for DSF biosynthesis, the cognate genes were deleted in both wild-type and rpfC deletion mutant backgrounds. Bioassay of these mutant strains did not reveal any effect on DSF biosynthesis (data not shown).
To verify specific binding of REC to RpfF, we conducted far Western blot analysis using two unrelated proteins, albicidin hydrolase (AlbD) (25) and the green fluorescent protein as negative controls (Fig. 6B). Western analysis showed that the anti-REC antiserum only recognized REC but not RpfF or the other two proteins (Fig. 6C). However, after soaking the membrane containing the separated proteins with a REC protein solution and subsequent stringent washes, probing with the anti-REC antiserum revealed a cross-reacting band at the position of the RpfF protein band in addition to the expected REC hybridization signal (Fig. 6D). The results thus demonstrated a specific REC-RpfF protein-protein interaction that is maintained throughout the stringent washing regime.

DISCUSSION
The hybrid sensor kinase RpfC of X. campestris pv. campestris acts to positively regulate the synthesis of extracellular enzymes and EPS virulence factors and negatively regulates the synthesis of the cell-cell signal DSF. The findings presented here are consistent with a model by which RpfC modulates FIGURE 5. Key amino acid residues implicated in REC-mediated repression of DSF biosynthesis. A, REC secondary structure prediction (GOR4 program) and the three key residues (indicated by vertical arrow) influencing REC activity. Arrows represent ␤-sheets, whereas cylinders represent ␣-helices. B, effect of key amino acid residue substitutions on DSF production. The coding sequence of minimum REC region R⌬8 (Fig. 4) and its alanine-substituted derivatives was cloned in pLAFR3 and expressed in rpfC deletion mutant Xc1⌬C. these diverse biological functions through two distinct molecular mechanisms; activation of virulence factor synthesis requires phosphorelay, whereas regulation of DSF production involves domain-specific protein-protein interaction with RpfF, the enzyme directing DSF synthesis.
Many microbial two-component systems with hybrid sensors, including AcrB/ArcA and TorS/TorR of E. coli (10,28), BvgS/BvgA of Bordetella pertussis (29), the LuxN(Q)-LuxU-LuxO of Vibrio harveyi and Vibrio cholerae (30), and the Kin/Spo of Saccharomyces cerevisiae (31), adopt a conserved four-step phosphorelay mechanism in signal transduction following signal perception. In most cases, His 1 -Asp 1 -His 2 is in the sensor and Asp 2 in the response regulator, although in Vibrio sp., His 2 is carried by the separate protein LuxO. RpfC contains several functional domains, i.e. transmembrane, HK, REC, and HPT with the three essential phosphorelay residues His 1 198 , Asp 1 512 , and His 2 657 located in the HK, REC, and HPT domains, respectively. Substitution of these three key residues with other amino acids or deletion of the HPT domain that contains the critical His 2 657 residue abrogated the RpfC activity in induction of EPS and virulence factor production. Given these results, we have concluded that RpfC uses the conserved His 1 -Asp 1 -His 2 phosphorelay mechanism to perceive and transduce environmental signals, which include DSF, with consequent activation of the synthesis of virulence factors.
Our findings also indicate that RpfC modulates DSF biosynthesis by a novel mechanism that is independent of the HPT domain and phosphorelay but involves interaction of the REC domain with RpfF, the key enzyme responsible for DSF synthesis (18). These conclusions depend upon sitedirected mutagenesis and deletion analysis of the chromosomal copy of the rpfC gene, examination of the effects of in trans expression of the REC domain and its truncated variants on DSF synthesis in the RpfC null mutant, and direct evidence of protein-protein interaction by co-immunoprecipitation and far Western analysis. Deletion analysis narrowed down the minimal region required for repression of DSF biosynthesis to a peptide of 107 amino acids. The alanine scanning mutagenesis peptide revealed that three amino acid replacements, Q496A, E504A, and I552A, partially decreased the repressor activity, although most alterations had no effect. Multiple mutations did not further reduce the repressor activity. Q496, E504, and I552 may either be directly involved in the interaction with the RpfF protein or in maintenance of a conformation of the REC domain that promotes that interaction.
Previous quantitative analysis has shown that DSF production in wild-type X. campestris pv. campestris is growth phasedependent and is maximal in the late stationary phase (20). In the well characterized quorum sensing systems involving acyl homoserine lactones, signal production is autoregulated; genes within the luxI family, which encode for acyl homoserine lactone synthases, are inducible by acyl homoserine lactone signals (32,33). However, in the DSF quorum sensing system, the transcription of rpfF occurs throughout growth and is not influenced by DSF (18,21). Furthermore, although rpfC mutants produce highly elevated levels of DSF, this is accompanied by only modest (up to 2-fold) changes in the level of rpfF transcript (17). The elevated DSF production in RpfC null mutants is seen throughout growth, even in the early growth phase where the wild-type X. campestris pv. campestris signal is low or undetectable (20). These findings, together with the data from this study, suggest a model for the control of virulence factor synthesis and DSF auto-induction by RpfC (Fig. 7). At low cell density or in an unconfined environment, the extracellular concentration of DSF is below a threshold, and autophosphorylation of RpfC is not initiated. Unphosphorylated RpfC adopts a structure that allows binding of RpfF to the REC domain, thus inhibiting DSF synthesis, which remains at a basal level. When the cell density is high or when bacteria enter a confined environment, the level of extracellular DSF increases. Upon reaching a threshold level, DSF binding causes RpfC to autophosphorylate, which results in a conformational change allowing release of RpfF, thus increasing DSF biosynthesis and facilitating the four-step phosphorelay that activates RpfG and the down-stream DSF regulon (Fig. 6B). In this manner, auto-induction of DSF could be achieved without substantial elevation in rpfF gene transcription. We cannot exclude the possibility that DSF production may also be regulated by substrate availability, which in turn may be negatively influenced by RpfC. However, this putative role of RpfC would also have to be independent of the phosphorelay, because H198A, D512V, and H657A variants all support wild-type levels of DSF.
Although phosphorylation-induced conformational changes in sensor kinases remain to be investigated, such phenomena have been well documented among response regulators (2). The distinct mechanisms of RpfC in modulating different functions have thus presented a new dimension to the conventional two-component signaling paradigms, although it remains to be seen how widespread such mechanisms are.